Numerical study of influence of oblique plate length and cooling rate on solidification and macrosegregation of A356 aluminum alloy melt with experimental comparison

• Model incorporates solid motion, dendritic breakage, coherency and species transport.
• It ignores buoyancy, solidification shrinkage/expansion & nucleation/growth kinetics.
• Effects of plate length and cooling rate are studied to predict various flow fields.
• Plate length of 250 mm and heat transfer coefficient of 2000 W/m2-K ideal for slurry.
• Industry suitability to produce semisolid cast components of globular microstructure.

The present study describes about the preparation of semisolid metal (SSM) slurry by using an oblique plate. In this process, A356 aluminum alloy melt partially solidifies while flowing over a bottom cooled oblique plate. Melt flow inertia shears columnar dendrites formed on plate wall into equiaxed/fragmented grains by resulting semisolid slurry at plate exit. Effects of plate length providing required shear and cooling rate enabling necessary solidification are investigated. A 3-phase numerical model vis-à-vis transport of mass, momentum, energy and species is developed for prediction of velocity, temperature, macrosegregation and solid fraction. Model uses volume of fluid (VOF) for tracking-metal-air-interface and finite volume method (FVM) with enthalpy based phase change algorithm for tracking-solid-liquid-interface within the metal. Darcy model is used for porous mushy zone. Slurry variable viscosity is represented by Oldenburg model. Stokes model incorporates solid phase movement and gravity effect along the flow. Dendrite fragmentation is considered for generation of moving solid phase. Solid movement is handled by coherency point and characteristic diameter of moving grains. Model neglects nucleation and growth kinetics, solidification shrinkage/expansion and thermo-solutal buoyancy. Slurry solid fractions at plate exit are 16%, 22%, and 27% for plate lengths of 200 mm, 250 mm, and 300 mm, respectively. And, are 5%, 22%, and 27% for heat transfer coefficients of 1000 W/m2-K, 2000 W/m2-K and 2500 W/m2-K, respectively. Numerical predictions agree well with experimental results.

Figure optionsDownload as PowerPoint slide